Advanced Data Management Technologies

Transcription

1 ADMT 2014/15 Unit 15 J. Gamper 1/44 Advanced Data Management Technologies Unit 15 Introduction to NoSQL J. Gamper Free University of Bozen-Bolzano Faculty of Computer Science IDSE

2 ADMT 2014/15 Unit 15 J. Gamper 2/44 Outline 1 Motivation 2 NoSQL 3 Categories of NoSQL Datastores Key-Value Stores Column Stores Document Stores Graph Databases

3 Motivation ADMT 2014/15 Unit 15 J. Gamper 3/44 Outline 1 Motivation 2 NoSQL 3 Categories of NoSQL Datastores Key-Value Stores Column Stores Document Stores Graph Databases

4 Motivation ADMT 2014/15 Unit 15 J. Gamper 4/44 New Trends

5 Motivation ADMT 2014/15 Unit 15 J. Gamper 5/44 Big Data The Digital Age/1 IDC/EMC annual report The Diverse and Exploding Digital Universe : The worlds information is doubling every two years. In 2011 the world will create a staggering 1.8 zettabytes. By 2020 the world will generate 50 times the amount of information... while IT staff to manage it will grow less than 1.5 times. New information taming technologies such as deduplication, compression, and analysis tools are driving down the cost of creating, capturing, managing, and storing information to one-sixth the cost in 2011 in comparison to zettabyte = bytes = 1 bio. terabytes

6 Motivation ADMT 2014/15 Unit 15 J. Gamper 6/44 Big Data The Digital Age/2 The New York Stock Exchange generates about 1 terabyte of new trade data per day. Facebook hosts approximately 10 billion photos, taking up one petabyte of storage. Ancestry.com, the genealogy site, stores around 2.5 petabytes of data. The Large Hadron Collider near Geneva will produce about 15 petabytes of data per year. But even an might produce a lot of data.

7 Motivation ADMT 2014/15 Unit 15 J. Gamper 7/44 3 V s of Big Data More V s are coming up: Veracity: accuracy and quality of data is difficult to control Value: it is important to turn big data it into value...

8 Motivation ADMT 2014/15 Unit 15 J. Gamper 8/44 RDBMSs The predominant choice in storing data up until now. First formulated in 1969 by Codd We are using RDBMS everywhere! BUT, are RDBMSs good in managing todays data?

9 Motivation ADMT 2014/15 Unit 15 J. Gamper 9/44 The Death of RDBMS?

10 Motivation ADMT 2014/15 Unit 15 J. Gamper 10/44 What is Wrong with RDBMSs? Nothing is wrong. They are great... SQL provides a rich, declarative language Database enforce referential integrity ACID properties are guaranteed Well understood by developers and administrators Support by many different languages

11 Motivation ADMT 2014/15 Unit 15 J. Gamper 11/44 ACID Properites Atomicity all or nothing Consistency any transaction will take the DB from one consistent state to another with no broken contraints (referential integrity) Isolation other operations cannot access data that has been modified during a transaction that has not yet completed Durability ability to recover the committed transaction updates against any kind of systems failure

12 Motivation ADMT 2014/15 Unit 15 J. Gamper 12/44 But there are some Problems with RDBMSs Problem: Complex objects Object/relational impedance mismatch Complicated to map rich domain model Performance issues: many rows in many tables, many joins,... Problem: Schema evolution Adding attributes to an object have to add columns to a table Expensive for large tables Holding locks on the tables for long time Problem: Semi-structered data Relational schema does not easily handle semi-structured data Common solutions Name/Value table: poor performance Serializable as Blob: fewer joins but no query capabilities Problem: Relational is hard to scale ACID does not scale well Easy to scale reads, but hard to scale writes

13 Motivation ADMT 2014/15 Unit 15 J. Gamper 13/44 One Size does not Fit All! There is nothing wrong with RDBMSs, but one size does not fit all! Alternative tools are available, just use the right tool. The rise of NoSQL databases marks the end of the era of relational database dominance. But NoSQL databases will not become the new dominators. Relational will still be popular, and used in the majority of situations. They, however, will no longer be the automatic choice.

14 NoSQL ADMT 2014/15 Unit 15 J. Gamper 14/44 Outline 1 Motivation 2 NoSQL 3 Categories of NoSQL Datastores Key-Value Stores Column Stores Document Stores Graph Databases

15 NoSQL ADMT 2014/15 Unit 15 J. Gamper 15/44 What is NoSQL? SQL or Not Only SQL or No to SQL? There is no standard definition! The term NoSQL was coined by Carlo Strozzi in 1998 In 2009 used by Eric Evans to refer to DBs which are non-relational, distributed and not conform to ACID. In 2009 first NoSQL conference Refers generally to data models that are non-relational, schema-free, non-(quite)-acid, horizontally scalable, distributed, easy replication support, simple API

16 NoSQL ADMT 2014/15 Unit 15 J. Gamper 16/44 Changing Requirements in the Web Age ACID properties are always desirable But, web applications have different needs from applications that RDBMS were designed for Low and predictable response time (latency) Scalability & elasticity (at low cost!) High availability Flexible schemas and semi-structured data Geographic distribution (multiple data centers) Web applications can (usually) do without Transactions, strong consistency, integrity Complex queries

17 NoSQL ADMT 2014/15 Unit 15 J. Gamper 17/44 CAP Theorem/1 Desired properties of web applications: Consistency the system is in a consistent state after an operation All clients see the same data Strong consistency (ACID) vs. eventual consistency (BASE) Availability the system is always on, no downtime Node failure tolerance all clients can find some available replica Software/hardware upgrade tolerance Partition tolerance the system continues to function even when split into disconnected subsets, e.g., due to network errors or addition/removal of nodes Not only for reads, but writes as well! CAP Theorem (E. Brewer, N. Lynch) In a shared-data system, at most 2 out of the 3 properties can be achieved at any given moment in time.

18 NoSQL ADMT 2014/15 Unit 15 J. Gamper 18/44 CAP Theorem/2 CA CP AP Single site clusters (easier to ensure all nodes are always in contact) e.g., 2PC When a partition occurs, the system blocks Some data may be inaccessible (availability sacrificed), but the rest is still consistent/accurate e.g., sharded database System is still available under partitioning, but some of the data returned my be inaccurate i.e., availability and partition tolerance are more important than strict consistency e.g., DNS, caches, Master/Slave replication Need some conflict resolution strategy

19 NoSQL ADMT 2014/15 Unit 15 J. Gamper 19/44 BASE Properties Requirements regarding reliability, availability, consistency and durability are changing. For a growing number of applications, availability and partition tolerance are more important than strict consistency. These properties are difficult to achieve with ACID properties The BASE properties forfeit the ACID properties of consistency and isolation in favor of availability, graceful degradation, and performance BASE properties Basically Available an application works basically all the time; Soft-state does not have to be consistent all the time; Eventual consistency but will be in some known state eventually. i.e., an application works basically all the time (basically available), does not have to be consistent all the time (soft-state) but will be in some known state eventually (eventual consistency

20 NoSQL ADMT 2014/15 Unit 15 J. Gamper 20/44 BASE vs. ACID Should be considered as a spectrum between the two extremes rather than two altenatives excluding each other ACID BASE Strong consistency Weak consistency stale data OK Isolation Availability first Focus on commit Best effort Nested transactions Approximate answers OK Availability? Aggressive (optimistic) Conservative (pessimistic) Simpler! Difficult evolution (e.g., schema) Faster Easier evolution

21 NoSQL ADMT 2014/15 Unit 15 J. Gamper 21/44 NoSQL Pros and Cons Advantages Massive scalability (horizontal scalability), i.e., machines can be added/removed High availability Lower cost (than competitive solutions at that scale) (Usually) Predictable elasticity Schema flexibility, sparse & semi-structured data Quicker and cheaper to set up Disadvantages Limited query capabilities (so far) Eventual consistency is not intuitive to program Makes client applications more complicated No standardization Portability might be an issue Insufficient access control

22 Categories of NoSQL Datastores ADMT 2014/15 Unit 15 J. Gamper 22/44 Outline 1 Motivation 2 NoSQL 3 Categories of NoSQL Datastores Key-Value Stores Column Stores Document Stores Graph Databases

23 Categories of NoSQL Datastores ADMT 2014/15 Unit 15 J. Gamper 23/44 Categories of NoSQL Datastores Key-Value stores Simple K/V lookups (DHT) Column stores Each key is associated with many attributes (columns) NoSQL column stores are actually hybrid row/column stores Different from pure relational column stores! Document stores Store semi-structured documents (JSON) Map/Reduce based materialisation, sorting, aggregation, etc. Graph databases Not exactly NoSQL... Cannot satisfy the requirements for high availability and scalability/elasticity very well.

24 Categories of NoSQL Datastores ADMT 2014/15 Unit 15 J. Gamper 24/44 Focus of Different NoSQL Data Models

25 Categories of NoSQL Datastores ADMT 2014/15 Unit 15 J. Gamper 25/44 Comparison of SQL and NoSQL Data Models Data Model Performance Scalability Flexibility Complexity Functionality Key-value Stores high high high none variable (none) Column Store high high moderate low minimal Document Store high variable (high) high low variable (low) Graph Database variable variable high high graph theory Relational Database variable variable low moderate relational algebra

26 Categories of NoSQL Datastores Key-Value Stores ADMT 2014/15 Unit 15 J. Gamper 26/44 Key-Value Stores Simple data model: global collection of key-value pairs. Favor high scalability to handle massive data over consistency Rich ad-hoc querying and analytics features are mostly omitted (especially joins and aggregate operations are set aside). Simple API with put and get Key-value stores have existed for a long time, e.g., Berkeley DB. Recent developments have been inspired by Distributed Hashtables and Amazon s Dynamo DeCandia et al., Dynamo: Amazon s Highly Available Key-value Store, SOSP 07 Another important free and open-source key-value store is Voldemort. Multiple types In memory: Memcache On disk: Redis, SimpleDB Eventually consistent: Dynamo, Voldemort

27 Categories of NoSQL Datastores Key-Value Stores ADMT 2014/15 Unit 15 J. Gamper 27/44 Dynamo P2P key-value store at Amazon, 2007 Context and requirements at Amazon Infrastructure: tens of thousands of servers and network components located in many data centers around the world Commodity hardware is used, where component failure is the standard mode of operation Amazon uses a highly decentralized, loosely coupled, service oriented architecture consisting of hundreds of services Low latency and high throughput Simple query model: unique keys, blobs, no schema, no multi-access Scale out (elasticity) Simple API get(key): returning a list of objects and a context put(key, context, object): no return value Key and object values are not interpreted but handled as an opaque array of bytes

28 Categories of NoSQL Datastores Key-Value Stores ADMT 2014/15 Unit 15 J. Gamper 28/44 Voldemort/1 Key-value store initially developed for and still used at LinkedIn Inspired by Amazon s Dynamo Features Written in Java Simple data model and only simple and efficient queries no joins or complex queries no constraints on foreign keys etc. Performance of queries can be predicted well P2P Scale-out / elastic Consistent hashing of keyspace Eventual consistency / high availability Pluggable storage BerkeleyDB, In Memory, MySQL

29 Categories of NoSQL Datastores Key-Value Stores ADMT 2014/15 Unit 15 J. Gamper 29/44 Voldemort/2 API consists of three functions: get(key): returning a value object put(key, value): writing an object/value delete(key): deleting an object Keys and values can be complex, compound objects as well consisting of lists and maps

30 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 30/44 Column Stores Data model: each key is associated with multiple attributes (i.e., columns) Hybrid row/column store Inspired by Google BigTable Examples: BigTable, HBase, Cassandra

31 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 31/44 BigTable BigTable at Google, 2006 A distributed storage system for managing structured data that is designed to scale to a very large size: petabytes of data across thousands of commodity servers Observation Key-value pairs are a useful building block, but should not be the only one Design goal: data model should be richer than simple key-value pairs, and support sparse semi-structured data, but simple enough that it lends itself to a very efficient flat-file representation

32 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 32/44 BigTable Data Model/1 Sparse, distributed, persistent multidimensional sorted map Values are stored as arrays of bytes (strings) which are not interpreted Values are addressed by (row, column, timestamp) dimensions Example: Multidimensional sorted map with information that a web crawler might emit Flexible number of rows representing domains Flexible number of columns first column contains the content of the web page the others store link text from referring domains Every value has a timestamp

33 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 33/44 BigTable Data Model/2 Row Keys are arbitrary strings Data is sorted by row key Tablet Row range is dynamically partitioned into tablets (sequence of rows) Range scans are very efficient Row keys should be chosen to improve locality of data access Column, Column Family Column keys are arbitrary strings, unlimited number of columns Column keys can be grouped into families Data in a CF is stored and compressed together (Locality Groups) Access control on the CF level

34 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 34/44 BigTable Data Model/3 Timestamps Each cell has multiple versions Can be manually assigned Versioning Automated garbage collection Retain last N versions or versions newer than TS Architecture Data stored on GFS 1 Master server Thousands of Tablet servers

35 Categories of NoSQL Datastores Column Stores ADMT 2014/15 Unit 15 J. Gamper 35/44 BigTable Architecture Data is stored in a 3-level hierarchy similar to B + -trees Chubby file contains location of root tablet Root tablet contains all tablet locations in Metadata table Metadata table stores locations of actual tablets

36 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 36/44 Document Stores Similar to a key-value database, but with a major difference: value is a document. Inspired by Lotus Notes Flexible schema Any number of fields can be added Document mainly stored in JSON or BSON formats Example document: { } day: [ 2010, 01, 23 ], products: { apple: { price: 10 quantity: 6 } kiwi: { price: 20 quantity: 2 } } checkout: 100

37 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 37/44 CouchDB/1 Schema-free, document store DB Documents stored in JSON format (XML in old versions) B-tree storage engine MVCC model, no locking No joins, no PK/FK (UUIDs are auto assigned) Implemented in Erlang 1st version in C++, 2nd in Erlang and 500 times more scalable Replication (incremental) Documents UUID Old versions retained Custom persistent views using MapReduce RESTful HTTP interface

38 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 38/44 CouchDB/2 Main abstraction and data structure is a document Consist of named fields that have a key/name and a value Field name must be unique in document Value may be a string, number, boolean, date, ordered list, map References to other documents (URIs, URLs) are possible but not checked by the DB Example document Title : CouchDB, Last editor : , Last modified : 9/23/2010, Categories : [ Database, NoSQL, Document Database ], Body : CouchDB is a..., Reviewed : false

39 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 39/44 MongoDB Document store DB written in C++ Full index support Replication & high availability Supports ad-hoc querying Fast in-place updates Officially supported drivers available for multiple languages C, C++, Java, Javascript, Perla, PHP, Python, Ruby Map/Reduce GridFS Commercial support

40 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 40/44 MongoDB/2 A database resides on a MongoDB server A MongoDB database consists of one or more collections of documents Schema-free, i.e., documents in a collection may be heterogeneous Main abstraction and data structure is a document Comparable to an XML document or a JSON document Documents are stored in BSON Similar to JSON, but binary representation for efficiency reasons Example document: { title : MongoDB, last editor : , last modified : new Date ( 9/23/2010 ), body : MongoDB is a..., categories : [ Database, NoSQL, Document Database ], reviewed : false }

41 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 41/44 MongoDB Example Create a collection named mycoll with 10,000,000 bytes of preallocated disk space and no automatically generated and indexed document-field db.createcollection( mycoll, size: , autoindexid: false) Add a document into mycoll db.mycoll.insert(title: MongoDB, last editor:... ) Retrieve a document from mycoll db.mycoll.find(categories: [ NoSQL, Document Databases ])

42 Categories of NoSQL Datastores Document Stores ADMT 2014/15 Unit 15 J. Gamper 42/44 MongoDB Deployment

43 Categories of NoSQL Datastores Graph Databases ADMT 2014/15 Unit 15 J. Gamper 43/44 Graph Databases Data Model Nodes Relations Properties Inspired by Euler s graph theory Examples: Neo4j, InfiniteGraph

44 ADMT 2014/15 Unit 15 J. Gamper 44/44 Summary New trends emerged in the past decade: big data, complexity, connectivity, diversity, etc. New requirements: consistency, availability and partitioning tolerance. NoSQL provides flexible solution for such requirements. NoSQL taxonomy Key-value stores Column stores Document stores Graph databases Use the right data model for the right problem.